| 01:03.27 | starseeker | wryly notes he needs to be sent to remedial OpenGL programming 101 for a week or so... |
| 01:56.54 | brlcad | just increased the performance of the the X dm interface by an order of magnitude |
| 01:58.51 | brlcad | refactors his proof-of-concept test code into something more production-worthy |
| 02:02.53 | CIA-73 | BRL-CAD: 03brlcad * r38832 10/brlcad/trunk/include/dm_xvars.h: add a helper macro INIT_XCOLOR for initializing X11 XColor structures. many locally scoped XColor's in use are initialized with random contents, which can leave flags set affecting behavior. |
| 02:14.58 | CIA-73 | BRL-CAD: 03brlcad * r38833 10/brlcad/trunk/BUGS: found a way to lock up mged reliably, if you Z or B the display while it's actively drawing a vlist. probably the geometry disappearing on the vlist, but easy to reproduce so should be traceable. |
| 02:15.57 | CIA-73 | BRL-CAD: 03brlcad * r38834 10/brlcad/trunk/src/libdm/color.c: initialize XColors to zero, requires headers and quellage due to memset. |
| 02:18.31 | CIA-73 | BRL-CAD: 03brlcad * r38835 10/brlcad/trunk/src/libdm/ (dm-X.c dm-tk.c): |
| 02:18.32 | CIA-73 | BRL-CAD: initialize all of the XColor local structures. otherwise, the structures can |
| 02:18.32 | CIA-73 | BRL-CAD: have random memory contents where the flags, padding, and initial pixel values |
| 02:18.32 | CIA-73 | BRL-CAD: might be non-zero. init to zero reflects expectation. also include debug code |
| 02:18.32 | CIA-73 | BRL-CAD: (ifdef0'd) on dm-X that shows a major performance problem. |
| 02:25.02 | CIA-73 | BRL-CAD: 03brlcad * r38836 10/brlcad/trunk/src/libdm/dm-X.c: reorganize to elimiate forward declarations, make get_color() HIDDEN |
| 02:57.15 | brlcad | woot |
| 02:59.16 | brlcad | okay, so I took a slight performance dip to not assume O(1) access and consume 200MB right off the bat, but now only use as much memory as the colors we draw.. about a 6x performance gain |
| 02:59.28 | brlcad | havoc is now nice and smooth |
| 03:07.51 | CIA-73 | BRL-CAD: 03brlcad * r38837 10/brlcad/trunk/src/libdm/dm-X.c: (log message trimmed) |
| 03:07.51 | CIA-73 | BRL-CAD: replace the proof-of-concept with something slightly less impressive but still |
| 03:07.51 | CIA-73 | BRL-CAD: with a dash of awesome. improve the performance of the X display manager by |
| 03:07.51 | CIA-73 | BRL-CAD: eliminating repeat XAllocColor() calls. original test bed showed about an order |
| 03:07.51 | CIA-73 | BRL-CAD: of magnitude improvement and had O(1) lookup time, but ate about 200MB of |
| 03:07.51 | CIA-73 | BRL-CAD: memory. this uses a bu_list to stash the colors so we only use as much as we |
| 03:07.52 | CIA-73 | BRL-CAD: need. seems to be still about a 5x-10x performance improvement over always |
| 03:08.43 | CIA-73 | BRL-CAD: 03brlcad * r38838 10/brlcad/trunk/TODO: need to release dm X color memory sooner rather than later, but definite improvement for release. |
| 03:12.38 | CIA-73 | BRL-CAD: 03brlcad * r38839 10/brlcad/trunk/NEWS: |
| 03:12.38 | CIA-73 | BRL-CAD: improved the performance of the X11 (dm X) wireframes in mged/archer/rtwizard by |
| 03:12.38 | CIA-73 | BRL-CAD: substantially eliminating memory allocations for colors. this is giving about a |
| 03:12.38 | CIA-73 | BRL-CAD: 5-10x performance boost to the wireframe allowing highly detailed models to |
| 03:12.38 | CIA-73 | BRL-CAD: display more interactively. |
| 03:19.19 | CIA-73 | BRL-CAD: 03brlcad * r38840 10/brlcad/trunk/include/conf/PATCH: the final stretch, bump minor to 7.16.8 for final steps and test pass. |
| 03:26.25 | CIA-73 | BRL-CAD: 03brlcad * r38841 10/brlcad/trunk/NEWS: |
| 03:26.26 | CIA-73 | BRL-CAD: The backout/-b option to nirt is suppose to back the ray origin point out of the |
| 03:26.26 | CIA-73 | BRL-CAD: geometry. Internally the origin point was actually being backed up by the |
| 03:26.26 | CIA-73 | BRL-CAD: bounding sphere diameter. This backout method will still miss geometry if your |
| 03:26.26 | CIA-73 | BRL-CAD: origin point is out past the bounding sphere distance. This option now backs out |
| 03:26.26 | CIA-73 | BRL-CAD: a bounding sphere radius distance in front of the bounding sphere. Fixed with |
| 03:26.27 | CIA-73 | BRL-CAD: r38575 by keith. |
| 03:29.56 | CIA-73 | BRL-CAD: 03brlcad * r38842 10/brlcad/trunk/NEWS: |
| 03:29.56 | CIA-73 | BRL-CAD: probably most significant user-visible change, bob fixed the Mac input bug that |
| 03:29.56 | CIA-73 | BRL-CAD: was affecting 10.5+ platforms where mouse input would stall during interactive |
| 03:29.56 | CIA-73 | BRL-CAD: rotations. this would occur after a zoom (mouse click) event. Bob removed the |
| 03:29.56 | CIA-73 | BRL-CAD: calls to focus for button presses (even though they should be fine, so there's |
| 03:29.56 | CIA-73 | BRL-CAD: still something screwy). |
| 03:30.56 | CIA-73 | BRL-CAD: 03brlcad * r38843 10/brlcad/trunk/NEWS: release should get tagged and bagged on the 28th. |
| 03:34.10 | CIA-73 | BRL-CAD: 03brlcad * r38844 10/brlcad/trunk/src/ (3 files in 3 dirs): bump batch files to .8 (this should NOT be necessary..shouldn't have or need version here) |
| 03:36.11 | CIA-73 | BRL-CAD: 03brlcad * r38845 10/brlcad/trunk/ChangeLog: update with changes from 2010-02-04 to present in preparation for release 7.16.8 |
| 03:36.50 | brlcad | ready to by sync'd to stable and final distcheck testing, alas I'm off to zzz now for a bit |
| 04:11.02 | CIA-73 | BRL-CAD: 03starseeker * r38846 10/brlcad/branches/STABLE/ (78 files in 37 dirs): Sync STABLE to r38845 |
| 04:21.05 | yukonbob | starseeker: is this sync a numbered release, or ?? |
| 04:22.31 | starseeker | yukonbob: we'll know tomorrow |
| 04:37.20 | yukonbob | starseeker: regression tests? |
| 07:18.50 | *** join/#brlcad Stattrav (~Stattrav@202.3.77.233) | |
| 07:37.30 | *** join/#brlcad Stattrav (~Stattrav@202.3.77.233) | |
| 08:27.13 | *** join/#brlcad Stattrav (~Stattrav@202.3.77.233) | |
| 11:34.44 | d-lo | Mernin all |
| 11:53.56 | CIA-73 | BRL-CAD: 03davidloman * r38847 10/rt^3/trunk/include/GS/GSCommon.h: Add LOCAL_DISCONNECT_REQ failure code. |
| 11:55.27 | CIA-73 | BRL-CAD: 03davidloman * r38848 10/rt^3/trunk/include/libNetwork/NetPortal.h: Add a default value to NetPortal::disconnectFromNetHost |
| 12:04.19 | *** join/#brlcad Stattrav (~Stattrav@202.3.77.233) | |
| 13:28.43 | ``Erik | shoulda noted the time when I started this msvc compile, it sure seems awful slow |
| 14:25.35 | CIA-73 | BRL-CAD: 03davidloman * r38849 10/rt^3/trunk/ (include/libNetwork/NetPortal.h src/libNetwork/NetPortal.cxx): Add isOpen to NetPortal for checking to see if the socket is still connected. |
| 14:29.49 | CIA-73 | BRL-CAD: 03davidloman * r38850 10/rt^3/trunk/tests/GS/GeometryServiceTest.cxx: Implement connected() in GeometryClient. Also notated the remaining functions in GeometryClient with //TODO tags. |
| 14:41.13 | ``Erik | hm, gqa seems to poop itself in regress, asplodes 'nuff where it forgets the argv[0] name, creates (unknown)-XXXX-bomb.log |
| 15:00.35 | *** join/#brlcad Stattrav (~Stattrav@202.3.77.233) | |
| 15:07.56 | ``Erik | (which distclean doesn't catch during distcheck) |
| 15:15.01 | brlcad | ``Erik: gqa -Av -v dumps out GB's of data .. if disk is filled, may crash out |
| 16:22.08 | CIA-73 | BRL-CAD: 03davidloman * r38851 10/rt^3/trunk/cmake/ProjectPrinter.cmake: Compact CMAKE printing into something more readable during config. |
| 16:23.42 | CIA-73 | BRL-CAD: 03bob1961 * r38852 10/brlcad/trunk/src/librt/primitives/bot/bot.c: Allocate resources for thickness if it's plate mode. |
| 17:11.30 | *** join/#brlcad Stattrav (~Stattrav@202.3.77.211) | |
| 17:19.48 | ``Erik | more than 124 gigs? |
| 17:23.56 | ``Erik | continues experimenting |
| 17:25.42 | CIA-73 | BRL-CAD: 03davidloman * r38853 10/rt^3/trunk/cmake/rt3commons.cmake: Implement CMAKE macros for enhancing/simplifying project building. New architecture maintains lists of INST/NOINST headers, source files, required libs, etc, and will support the ability to install rt3 onto a computer. |
| 17:30.39 | CIA-73 | BRL-CAD: 03bob1961 * r38854 10/brlcad/trunk/src/librt/primitives/bot/bot.c: If the new bot is plate mode, set the thickness and face_mode for each face. |
| 17:48.54 | CIA-73 | BRL-CAD: 03davidloman * r38855 10/rt^3/trunk/cmake/rt3commons.cmake: Fix some fouled INCLUDE path logic. Projects should now actually include all the paths set for it to include. |
| 17:52.42 | CIA-73 | BRL-CAD: 03davidloman * r38856 10/rt^3/trunk/src/GE/CMakeLists.txt: Remove libGeometry, libImage, libNumeric, libRaytrace from build system for now. |
| 18:03.55 | ``Erik | wonders if gqa grid resolutions should be changed to a bigger edge length for regression suite duration... is getting tired of waiting during distcheck |
| 18:31.36 | CIA-73 | BRL-CAD: 03davidloman * r38857 10/rt^3/trunk/ (CMakeLists.txt cmake/rt3commons.cmake): Make GLOBAL_INCLUDE_DIRS cache in cmake. Add a project's source path to the GLOBAL_INCLUDE_DIRS var if the project is a library. |
| 18:31.49 | CIA-73 | BRL-CAD: 03davidloman * r38858 10/rt^3/trunk/TODO: Modify TODO |
| 18:36.17 | CIA-73 | BRL-CAD: 03davidloman * r38859 10/rt^3/trunk/ (60 files in 15 dirs): Modify libUtility cmake to use newer build macros. Cascading changes a plenty. |
| 18:43.08 | CIA-73 | BRL-CAD: 03davidloman * r38860 10/rt^3/trunk/cmake/rt3commons.cmake: CMAKE library header generation macro was failing to put quotes around header file names. Fixed. |
| 18:45.42 | CIA-73 | BRL-CAD: 03davidloman * r38861 10/rt^3/trunk/ (6 files in 2 dirs): Refactor vague Utils to more precise DataStreamUtils |
| 18:55.12 | CIA-73 | BRL-CAD: 03davidloman * r38862 10/rt^3/trunk/ (9 files in 5 dirs): Enforce the use of libUtility.h rather than the individual header files (for now) |
| 19:03.30 | CIA-73 | BRL-CAD: 03davidloman * r38863 10/rt^3/trunk/ (13 files in 7 dirs): Modify RT3_PROJECT_BUILD_EXE() and RT3_PROJECT_BUILD_LIB() to no longer need a target lib.exe name. Uses the PROJECT_NAME instead. |
| 19:08.32 | CIA-73 | BRL-CAD: 03davidloman * r38864 10/rt^3/trunk/ (CMakeLists.txt cmake/rt3commons.cmake): Add VERBOSE_CMAKE_OUTPUT flag for quelling cmake screen vomit. |
| 19:09.51 | *** join/#brlcad Stattrav (~Stattrav@202.3.77.211) | |
| 19:22.12 | CIA-73 | BRL-CAD: 03brlcad * r38865 10/brlcad/trunk/NEWS: include verbose wordage summary for release 7.16.8 detailing NURBS; MGED stability, performance, and Windows 64-bit; and Archer work. |
| 20:11.32 | *** join/#brlcad R0b0t1 (~Enigma@unaffiliated/r0b0t1) | |
| 20:13.39 | CIA-73 | BRL-CAD: 03davidloman * r38866 10/rt^3/trunk/cmake/rt3commons.cmake: Fix some logic concerning MOC-ing of qt4 classes. |
| 20:20.26 | CIA-73 | BRL-CAD: 03brlcad * r38867 10/brlcad/branches/STABLE/ (NEWS src/librt/primitives/bot/bot.c): merge trunk to STABLE from r38845 to HEAD r |
| 20:24.21 | brlcad | well that didn't work so well |
| 20:25.40 | CIA-73 | BRL-CAD: 03davidloman * r38868 10/rt^3/trunk/ (106 files in 10 dirs): Converted over libGS, geoserv and libNetwork to use the newer cmake macros. Stopped halfway through, commiting what I have. Will continue asap. Compile is currently broken. |
| 20:27.22 | CIA-73 | BRL-CAD: 03erikgreenwald * r38869 10/brlcad/trunk/ (TODO regress/gqa.sh): Tune back the grid resolution for the gqa regression test and note that the plot file issue should be looked at. |
| 21:02.44 | CIA-73 | BRL-CAD: 03brlcad * r38870 10/brlcad/branches/STABLE/ (TODO regress/gqa.sh): merge trunk to STABLE from r38867 to HEAD r38869 |
| 21:04.41 | *** join/#brlcad Stattrav (~Stattrav@202.3.77.233) | |
| 21:21.56 | CIA-73 | BRL-CAD: 03brlcad * r38871 10/brlcad/trunk/src/lgt/ (glob.c prnt.c): remove the pointless version global |
| 21:25.08 | CIA-73 | BRL-CAD: 03brlcad * r38872 10/brlcad/branches/STABLE/src/libdm/dm-rtgl.c: somehow, this file is out of sync with head revision. manually sync. |
| 21:35.03 | CIA-73 | BRL-CAD: 03brlcad * r38873 10/brlcad/trunk/ (6 files in 5 dirs): remove unnecessary rcs variables that just make merges potentially difficult and source tree diff comparisons more complicated to review. |
| 22:10.15 | CIA-73 | BRL-CAD: 03erikgreenwald * r38874 10/isst/trunk/src/gui.c: start setting things up to automatically set lookat, location, and mouse speed |
| 22:44.48 | CIA-73 | BRL-CAD: 03r_weiss * r38875 10/brlcad/trunk/src/conv/obj-g_new.c: adding command line option logic, nmg creation testing |
| 23:56.39 | ``Erik | neat, got my joystick working with flightgear on my mac |
| 23:56.42 | ``Erik | kinda |